FIG. 1 shows a liquid droplet 4 in contact with a solid surface 2 and in static equilibrium. The contact angle θ6 is defined as shown in FIG. 1, and is determined by the balancing of the surface tension components between the solid-liquid (YSL 8), liquid-gas (YLG 10) and solid gas (YSG 12) interfaces, as shown, such that:
                              cos          ⁢                                          ⁢          θ                =                                            γ              SG                        -                          γ              SL                                            γ            LG                                              (                  equation          ⁢                                          ⁢          1                )            
The contact angle θ is thus a measure of the hydrophobicity of the surface. Surfaces may be described as hydrophilic if θ<90 degrees or hydrophobic if θ>90 degrees, and as more or less hydrophobic/hydrophilic according to the difference between the contact angle and 90 degrees. FIG. 2 shows a liquid droplet 4 in static equilibrium on hydrophilic 14 and hydrophobic 16 material surfaces with respective contact angles θ6.
FIG. 3 shows the case where a droplet straddles two regions of different hydrophobicity (e.g., the hydrophobic surface 16 and the hydrophilic surface 14). In this case the situation is non-equilibrium and in order to minimise the potential energy the droplet will move laterally towards the region of greater hydrophilicity. The direction of motion is shown as 18.
If the droplet consists of an ionic material, it is well known that it is possible to change the hydrophobicity of the surface by the application of an electric field. This phenomenon is termed electrowetting. One means for implementing this is using the method of electrowetting on dielectric (EWOD), shown in FIG. 4.
A lower substrate 25 has disposed upon it a conductive electrode 22, with an insulator layer 20 deposited on top of that. The insulator layer 20 separates the conductive electrode 22 from the hydrophobic surface 16 upon which the droplet 4 sits. By applying a voltage V to the conductive electrode 22, the contact angle θ 6 can be adjusted. An advantage of manipulating contact angle θ 6 by means of EWOD is that the power consumed is low, being just that associated with charging and discharging the capacitance of the insulator layer 20.
FIG. 5 shows an alternative and improved arrangement whereby a top substrate (counter-substrate) 36 is also supplied, containing an electrode 28 coated with a hydrophobic layer 26. A voltage V2 may be applied to the electrode 28 such that the electric field at the interfaces of the liquid droplet 4 and hydrophobic layer 26 and substrate 16 is a function of the difference in potential between V2 and V. A spacer 32 may be used to fix the height of the channel layer in which the droplet 4 is constrained. In some implementations the channel volume around the droplet 4 may be filled by a non-ionic liquid, e.g. oil 34. The arrangement of FIG. 5 is advantageous compared to that of FIG. 1 for two reasons: Firstly it is possible to generate larger and better controlled electric fields at the surfaces where the liquid droplet contacts the hydrophobic layer. Secondly the liquid droplet is sealed within the device, preventing loss due to evaporation etc.
The above background art is all well known and a more detailed description can be found in standard textbooks, e.g. “Introduction to Microfluidics”, Patrick Tabeling, Oxford University Press, ISBN 0-19-856864-9, section 2.8.
U.S. Pat. No. 6,565,727 (Shenderov, issued May 20, 2003) discloses a passive matrix EWOD device for moving droplets through an array. The device is constructed as shown in FIG. 6. The conductive electrode of the lower substrate 25 is patterned so that a plurality of electrodes 38 (e.g., 38A and 38B) are realised. These may be termed the EW drive elements. The term EW drive element may be taken in what follows to refer both to the electrode 38 associated with a particular array element, and also to the node of an electrical circuit directly connected to this electrode 38. By applying different voltages, termed the EW drive voltages, (e.g. V and V3) to different electrodes (e.g. drive elements 38A and 38B), the hydrophobicity of the surface can be controlled, thus enabling droplet movement to be controlled.
U.S. Pat. No. 6,911,132 (Pamula et al, issued Jun. 28, 2005) discloses an arrangement, shown in FIG. 7, whereby the conductive layer 22 on the lower substrate 25 is patterned to form a two dimensional array 42. By the application of time dependent voltage pulses to some or all of the different drive elements it is thus possible to move a liquid droplet 4 though the array on a path 44 that is determined by the sequence of the voltage pulses. U.S. Pat. No. 6,565,727 further discloses methods for other droplet operations including the splitting and merging of droplets and this mixing together of droplets of different materials. In general the voltages required to perform typical droplet operations are relatively high. Values in the range 20-60V are quoted in prior art (e.g. U.S. Pat. No. 7,329,545 (Pamula et al., issued Feb. 12, 2008), Lab on a Chip, 2002, Vol. 2, pages 96-101). The value required depends principally on the technology used to create the insulator and hydrophobic layers.
U.S. Pat. No. 7,255,780 (Shenderov, issued Aug. 14, 2007) similarly discloses a passive matrix EWOD device used for carrying out a chemical or biochemical reaction by combining droplets of different chemical constituents.
It may be noted that it is also possible, albeit generally not preferred, to implement an EWOD system to transport droplets of oil immersed in an aqueous ionic medium. The principles of operation are very similar to as already described, with the exception that the oil droplet is attracted to the regions where the conductive electrode is held at low potential.
When performing droplet operations it is in general very useful to have some means of sensing droplet position, size and constitution. This can be implemented by a number of means. For example an optical means of sensing may be implemented by observing droplet positions using a microscope. A method of optical detection using LEDs and photo-sensors attached to the EWOD substrate is described in Lab Chip, 2004, 4,310-315.
One particularly useful method of sensing is measuring the electrical impedance between an electrode 38 of the lower (patterned) conductive electrode 22 and the electrode 28 of the top substrate. FIG. 8 shows an approximate circuit representation 52 of the impedance in the case where a droplet 4 is present. A capacitor 46 representing the capacitance Ci of the any insulator layers (including the hydrophobic layers) is in series with the impedance of the droplet 4 which can be modelled as a resistor 50 with resistance Rdrop in parallel with a capacitor 48 with capacitance Cdrop. FIG. 9 shows the corresponding circuit representation 56 in the case where there is no droplet present. In this instance the impedance is that of the insulator layer capacitor 46 in series with a capacitor 54 representing the capacitance Cgap of the cell gap. Since the overall impedance of this arrangement has no real (i.e. resistive) component, the total impedance can be represented as a frequency dependent capacitor of value CL.
FIG. 10 shows schematically the dependence of CL with frequency in the cases where a droplet 4 is present (represented by dashed line 52) and where a droplet 4 is absent (represented by solid line 56). It can thus be readily appreciated that by measuring the impedance it is possible to determine whether or not a droplet 4 is present at a given node. Furthermore the value of the parameters Cdrop and Rdrop are a function of the size of the droplet 4 and the conductivity of the droplet 4. It is therefore possible to determine information relating to droplet size and droplet constitution by means of a measurement of capacitance. Sensors and Actuators B, Vol. 98 (2004) pages 319-327 describes a method for measuring droplet impedance by connecting external PCB electronics to an electrode in an EWOD array. However a disadvantage of this method is that the number of array elements at which impedance can be sensed is limited by the number of connections that can be supplied to the device. Furthermore this is not an integrated solution with external sensor electronics being required. The paper also describes how measured impedance can be used to meter the size of droplets and how droplet metering can be used to accurately control the quantities of reagents of chemical or biochemical reactions performed using an EWOD device. Impedance measurements at one or more locations could also be used for any of the following:                Monitor the position of droplets within an array        Determining the position of droplets within the array as a means of verifying the correct implementation of any of the previously droplet operations        Measuring droplet impedance to determine information regarding drop constitution, e.g. conductivity.        Measuring droplet impedance characteristics to detect or quantify a chemical or biochemical reaction.        
EWOD devices have been identified as a promising platform for Lab-on-a-chip (LoaC) technology. LoaC technology is concerned with devices which seek to integrate a number of chemical or biochemical laboratory functions onto a single microscopic device. There exists a broad range of potential applications of this technology in areas such as healthcare, energy and material synthesis. Examples include bodily fluid analysis for point-of-care diagnostics, drug synthesis, proteomics, etc.
A complete LoaC system could be formed, for example, by an EWOD device to other equipment, for example a central processing unit (CPU) which could be configured to perform one or more multiple functions, for example:                Supply voltage and timing signals to the AM-EWOD        Analyse sensor data returned from the AM-EWOD        Store in memory programmed data and/or sensor data        Perform sensor calibration operations upon demand and store sensor calibration information in memory        Process sensor data received from the AM-EWOD, including making adjustments based on saved calibration data        Adjust and control the voltage levels and timings of sensor control signals        Send digital or analogue data to the AM-EWOD for implementing droplet operations        Send digital or analogue data to the AM-EWOD for implementing droplet operations whose content depends on measured sensor output data        Adjust the voltage levels of the signals written to the EW drive electrodes in accordance with measured sensor output data.        
Thin film electronics based on thin film transistors (TFTs) is a very well known technology which can be used, for example, in controlling Liquid Crystal (LC) displays. TFTs can be used to switch and hold a voltage onto a node using the standard display pixel circuit shown in FIG. 11. The pixel circuit consists of a switch transistor 68, and a storage capacitor 57. By application of voltage pulses to the source addressing line 62 and gate addressing line 64, a voltage Vwrite can be written to the write node 66 and stored in the pixel. By applying a different voltage to the electrode of the counter-substrate CP 70, a voltage is thus maintained across the liquid crystal capacitance 60 within the pixel.
Many modern displays use an Active Matrix (AM) arrangement whereby a switch transistor is provided in each pixel of the display. Such displays often also incorporate integrated driver circuits to supply voltage pulses to the row and column lines (and thus program voltages to the pixels in an array). These are realised in thin film electronics and integrated onto the TFT substrate. Circuit designs for integrated display driver circuits are very well known. Further details on TFTs, display driver circuits and LC displays can be found in standard textbook, for example “Introduction to Flat Panel Displays”, (Wiley Series in Display Technology, WileyBlackwell, ISBN 0470516933).
U.S. Pat. No. 7,163,612 (Sterling et al., issued Jan. 16, 2007) describes how TFT-based electronics may be used to control the addressing of voltage pulses to an EWOD array using circuit arrangements very similar to those employed in AM display technologies. FIG. 12 shows the approach taken. In contrast with the EWOD device shown in FIG. 6, the lower substrate 25 is replaced by a TFT substrate 72 having thin film electronics 74 disposed upon it. The thin film electronics 74 are used to selectively program voltages to the patterned conductive layer 22 used for controlling electrowetting. It is apparent that the thin film electronics 74 can be realised by a number of well known processing technologies, for example silicon-on-insulator (SOI), amorphous silicon on glass or low temperature polycrystalline silicon (LTPS) on glass.
Such an approach may be termed “Active Matrix Electrowetting on Dielectric” (AM-EWOD). There are several advantages in using TFT-based electronics to control an EWOD array, namely:                Driver circuits can be integrated onto the AM-EWOD substrate. An example arrangement is shown in FIG. 13. Control of the EWOD array 42 is implemented by means of integrated row driver 76 and column driver 78 circuits. A serial interface 80 may also be provided to process a serial input data stream and write the required voltages to the array 42. The number of connecting wires 82 between the TFT substrate 72 (FIG. 12) and external drive electronics, power supplies etc. can be made relatively few, even for large array sizes.        TFT-based electronics are well suited to the AM-EWOD application. They are cheap to produce so that relatively large substrate areas can be produced at relatively low cost.        It is possible to incorporate TFT-based sensing into Active Matrix controlled arrays. For example US20080085559 (Hartzell et al., published Apr. 10, 2008) describes a TFT based active matrix bio-sensor utilising cantilever based arrays.        
A further advantage of using TFT based electronics to control an AM-EWOD array is that, in general, TFTs can be designed to operate at much higher voltages than transistors fabricated in standard CMOS processes. However the large AM-EWOD programming voltages (20-60V) can in some instances still exceed the maximum voltage ratings of TFTs fabricated in standard display manufacturing processes. To some extent it is possible to modify the TFT design to be compatible with operation at higher voltages, for example by increasing the device length and/or adding Gate-Overlap-Drain (GOLD) or Lightly Doped Drain (LDD) structures. These are standard techniques for improving Metal-On-Semiconductor (MOS) device reliability which can be found described, for example, in “Hot Carrier Effects in MOS Devices”, Takeda, Academic Press Inc., ISBN 0-12-682240-9, pages 40-42. However such modifications to device design may impair the TFT performance. For example, structural modifications to improve reliability may increase device self resistance and inter-terminal capacitances. The effects of this are particularly deleterious for devices which are required to operate at high speed or to perform analogue circuit functions. It is therefore desirable to restrict the use of modified high voltage devices to only those functions for which a high voltage capability is necessary, and to design driver circuits such that as few devices as possible are required to operate at the highest voltages.
Fluid manipulation by means of electrowetting is also a well known technique for realizing a display. Electronic circuits similar or identical to those used in conventional Liquid Crystal Displays (LCDs) may be used to write a voltage to an array of EW drive electrodes. Coloured droplets of liquid are located at the EW drive electrodes and move according the programmed EW drive voltage. This in turn influences the transmission of light through the structure such that the whole structure functions as a display. An overview of electrowetting display technology can be found in “Invited Paper: Electro-wetting Based Information Displays”, Robert A . Hayes, SID 08 Digest pp 651-654.
In recent years there has been much interest in realising AM displays with an array based sensor function. Such devices can be used, for example as user input devices, e.g. for touch-screen applications. One such method for user interaction is described in US20060017710 (Lee et al., published Jan. 26, 2006) and shown in FIG. 14. When the surface of the device is touched, for example by means of a fingertip or a stylus 90, the liquid crystal layer 92 is compressed in the vicinity of the touch. Integrated thin film electronics 74 disposed on the TFT substrate 72 can be used to measure the change in capacitance 60 of the LC layer and thus measure the presence 84 or absence 86 of touch. If the thin film electronics 74 are of sufficient sensitivity it is also possible to measure the pressure with which the surface is touched.
U.S. Pat. No. 7,163,612 noted above also describes how TFT-based sensor circuits may be used with an AM-EWOD, e.g. to determine drop position. In the arrangement described there are two TFT substrates, the lower one being used to control the EWOD voltages, and the top substrate being used to perform a sensor function.
A number of TFT based circuit techniques for writing a voltage to a display pixel and measuring the capacitance at the pixel are known. US20060017710 discloses one such an arrangement. The circuit is arranged in two parts which are not directly connected electrically, shown FIG. 15. The operation of the voltage write portion 101 of the pixel circuit is identical to a standard display pixel circuit as has already been described in relation to FIG. 11. The operation of the sensor portion 103 of the pixel circuit is as now described. For the sensor array row being sensed, a voltage pulse is supplied to a sensor row select line RWS 104. The potential of the sense node Vsense 102 will then increase by an amount that depends on the relative values of the LC capacitance CLC2 100 and the fixed reference capacitor CS 98 (and also on parasitic capacitances including those associated with the transistor 94). The potential of the sense node 102 can be measured as follows. Transistor 94 in combination with a load device (not shown) acts as standard source follower arrangement as is very well known, e.g. “CMOS Analog Circuit Design”, Allen and Holberg, ISBN-10: 0195116441, section 5.3. Since the value of the capacitor CS 98 is known, measurement of column output voltage at the sensor output line COL 106 is thus a measure of the LC capacitance. A notable feature of the whole arrangement is that the write node 66 and the sense node 102 are not electrically connected. Direct connection is not necessary or desirable since detection of touch does not require the LC capacitance of the entire pixel to be measured, but instead only the capacitance of a sample portion of it.
A disadvantage of the above circuit is that there is no provision of any DC current path to the sense node 102. As a result the potential of this node may be subject to large pixel-to-pixel variations, since fixed charge at this node created during the manufacturing process may be variable from pixel-to-pixel. An improvement to this circuit is shown in FIG. 16. Here an additional diode 110 is connected to the sense node 102. The potential at the anode of the diode RST 108 is maintained such that the diode 110 is reversed biased. This potential may be taken high to forward bias the diode 110 for a brief time period before the voltage pulse is applied to the sensor row select line 104. The effect of the voltage pulse applied to reset line RST 108 is to reset the potential of the sense node 102 to an initial value which can be very well controlled. This circuit arrangement therefore has the advantage of reduced pixel-pixel variability in the measured output voltage.
In general it may be noted that in this application, both the value of the LC capacitance and the change in capacitance associated with touch are very small (of order a few fF). One consequence of this is that reference capacitor CS 98 can also be made very small (typically a few fF). The small LC capacitance also makes changes difficult to sense. British applications GB 0919260.0 (Brown, published May 11, 2011) and GB 0919261.8 (Brown, published May 11, 2011) describe means of in-pixel amplification of the small signals sensed. However in an EWOD device the capacitances presented by droplets are much larger and amplification is generally not required.
As well as implementing sensor pixel circuits onto a TFT substrate it is also well known to integrate sensor driver circuits and output amplifiers for the readout of sensor data onto the same TFT substrate, as described for example for an imager-display in “A Continuous Grain Silicon System LCD with Optical Input Function”, Brown et al. IEEE Journal of Solid State Circuits, Vol. 42, Issue 12, December 2007 pp 2904-2912. The same reference also describes how calibration operations may be performed to remove fixed pattern noise from the sensor output.
There are several methods that may be used to form a capacitor circuit element in a thin film manufacturing process as would be used for example to manufacture a display. Capacitors can be formed for example using the source and gate metal layers as the plates, these layers being separated by an interlayer dielectric. In situations where it is important to keep the physical layout footprint of the capacitor it is often convenient to use a metal-oxide-semiconductor (MOS) capacitor as described in standard textbooks, e.g. Semiconductor Device Modelling for VLSI, Lee et al., Prentice-Hall, ISBN 0-13-805656-0, pages 191-193. A disadvantage of MOS capacitors is that the capacitance becomes a function of the terminal biases if the potentials are not arranged so that the channel semiconductor material is completely in accumulation. FIG. 17 shows at 124 the typical characteristics of a MOS capacitor 120 where the semiconductor material 122 is doped n-type. Plate A of the MOS capacitor 120 is formed by a conductive material (e.g. the gate metal) and plate B is the n-doped semiconductor material 122. The capacitance is shown in dotted line 126 as a function of the difference in voltage (bias voltage VAB) between the two plates A and B. Above a certain bias voltage Vth corresponding to approximately the threshold voltage of the n-type doped semiconductor material 122, the semiconductor material 122 is in accumulation and the capacitance is large and independent of voltage. If VAB is less than Vth the capacitance becomes smaller and voltage dependent as the n-type semiconductor material 122 becomes depleted of charge carriers.
FIG. 18 at 130 shows the corresponding situation where in this case the semiconductor material 128 forming plate B of the MOS capacitor 120 is doped p-type. In this case the maximum capacitance is obtained when VAB is below the threshold voltage Vth and the channel semiconductor material 128 is in accumulation.
A known lateral device type which can be realised in thin film processes is a gated P-I-N diode 144, shown FIG. 19. The gated P-I-N diode is formed from a layer of semiconductor material consisting of a p+ doped region 132, a lightly doped region 134 which may be either n-type or p-type, and an n-+ region 136. Electrical connections, e.g. with metal, are made to the p+ and n+ regions (132 and 136) to respectively form the anode terminal 137 and cathode terminal 138 of the device 144. An electrically insulating layer 142 is disposed over some or all of the lightly doped region 134, and a conductive layer forms the third gate terminal 140 of the device 144 denoted the gate terminal. Further description and explanation of the operation of such a device can be found in “High performance gated lateral polysilicon PIN diodes”, Stewart and Hatalis, Solid State Electronics, Vol. 44, Issue 9, p 1613-1619. FIG. 20 shows a circuit symbol which may be used to represent the gated P-I-N diode 144 and the three connecting terminals 137, 138 and 140 corresponding to the anode, cathode, and gate, respectively.
The gated P-I-N diode 144 may be configured as a type of MOS capacitor by connecting the anode and cathode terminals together to form one terminal of the capacitor, and by using the gate terminal 140 to form the other terminal.
By connecting the gated P-I-N diode 144 in this way it functions in a similar way to the MOS capacitor as already described, with the important difference that most of the channel region remains accumulated with carriers almost regardless of the voltage between the terminals. The operation of the gated P-I-N diode 144 connected in this way is illustrated in FIG. 21. In the case represented at 158 where the voltage potential VA 157 supplied to the gate terminal 140 exceeds the voltage potential VB 155 applied to the anode terminal 137 and cathode terminal 138 (plus the channel material threshold voltage), the majority of the channel 160 (the lightly doped region 134 in FIG. 19) becomes accumulated with negatively charged carriers (electrons) supplied from the cathode terminal 138 of the gated P-I-N diode 144. The capacitance between the gate terminal 140 and the (connected together) anode terminal 137 and cathode terminal 138 then approximates to that of a MOS capacitor in accumulation. Similarly, in the case represented at 162 where VA<VB, the majority of the channel 160 becomes accumulated with positive charge carriers (holes) supplied from the anode terminal 137 of the gated P-I-N diode 144. The capacitance between the gate terminal 140 and the anode/cathode terminals 137/138 again approximates to that of a MOS capacitor in accumulation. It is also possible to form a voltage dependent capacitor from a gated P-I-N diode 144, by connecting a bias voltage to the anode terminal 137 of the device relative to the cathode terminal 138. The bias applied, −VX, should be chosen such that the gated P-I-N diode 144 remains reverse biased.
The dashed line 164, 166, 168 in FIG. 22 shows schematically the capacitance versus voltage behaviour of the gated P-I-N diode 144 when connected as shown in FIG. 21. It can be seen that at both positive 164 and negative 166 bias voltages VAB (where VAB=VA−VB), the gated P-I-N diode 144 behaves like a MOS capacitor in accumulation. A small dip in the capacitance 168 appears as indicated around the threshold voltage of the material within the channel 160 (region 134 in FIG. 19). In the case represented by dotted line 176, a bias voltage−VX is applied to the anode terminal 137 relative to the cathode terminal 138. As is shown, the manner in which the capacitance varies as a function of the voltage difference between the anode terminal and the cathode terminal may be modified with application of the bias voltage−VX.
In both AM-EWOD and AM displays a number of possible alternative configurations for storing a programmed write voltage within a pixel are possible. For example a static random-access memory (SRAM) cell can be used to store the programmed voltage as is shown in FIG. 23. The SRAM cell 194 has CK and CKB clock inputs, and data input IN and a data output OUT. CK and CKB are connected to signals that are logical complements. Data is read into the cell via the IN input and transistor 290 when the CK input is high and the CKB input low; the data is passed through the two inverters 294 and 296 and presented at the output OUT. When CK is subsequently set low and CKB high transistor 292 closes a bi-stable loop such that the two inverters 294 and 296 retain the data.
An alternative technology for implementing droplet microfluidics is dielectrophoresis. Dielectrophoresis is a phenomenon whereby a force may be exerted on a dielectric particle by subjecting it to a varying electric field. An introduction may be found in “Introduction to Microfluidics”, Patrick Tabeling, Oxford University Press (January 2006), ISBN 0-19-856864-9, pages 211-214. “Integrated circuit/microfluidic chip to programmably trap and move cells and droplets with dielectrophoresis”, Thomas P Hunt et al, Lab Chip, 2008,8,81-87 describes a silicon integrated circuit (IC) backplane to drive a dielectropheresis array for digital microfluidics. This reference also includes an array-based integrated circuit for supplying drive waveforms to array elements.